Solana Token Early Caller Finder

Find the influential X (Twitter) accounts that posted a Solana token's contract address before or during its first price pump — then track them across multiple tokens to identify consistent early callers.

How It Works

   CA input
      │
      ▼
┌─────────────┐     ┌──────────────┐     ┌───────────┐     ┌─────────┐     ┌───────────┐
│ DexScreener │────▶│  X / Twitter │────▶│   Score   │────▶│  CSV +  │────▶│ Watchlist │
│  pump time  │     │    search    │     │  & rank   │     │ console │     │  update   │
└─────────────┘     └──────────────┘     └───────────┘     └─────────┘     └───────────┘

Pump timestamp — Queries the DexScreener API for pair creation time; estimates peak volume at +15 min.
X search — Uses twikit (with automatic patches for X's latest auth changes) to search for tweets containing the CA. Searches both "Latest" and "Top" feeds, up to 200 tweets.
Scoring — Ranks by influence: (followers × 0.4) + (likes × 0.4) + (retweets × 0.2), filtered to pre-peak tweets.
Output — Saves ranked results to CSV and prints the top callers.
Watchlist — Merges results into watchlist.json, building a profile of each caller across every token you investigate.

Requirements

Python 3.10+
An X (Twitter) account

Installation

git clone https://github.com/kerslakechris/test.git
cd test
pip install -r requirements.txt

Dependencies

Package	Purpose
`twikit`	X/Twitter scraping via GraphQL
`aiohttp`	Async HTTP for DexScreener API
`python-dotenv`	Load `.env` credentials
`streamlit`	Web-based GUI
`pandas`	Data tables in the GUI

Setup — X Authentication

You need an authenticated X session stored in cookies.json. There are two ways to set this up:

Option A — Browser cookies (recommended)

X frequently blocks programmatic login. Importing cookies from your browser is the most reliable method.

Log in to x.com in your browser
Open DevTools (F12) → Application → Cookies → https://x.com
Run the setup command:

python solana_callers.py setup-cookies

Paste the cookie values when prompted:
- auth_token — required
- ct0 — required
- kdt, twid, guest_id — optional but recommended

Cookies expire after ~24 hours. Re-run setup-cookies when you get a 401 error.

Option B — Credentials via .env

Create a .env file from the template:

cp .env.example .env

Then edit .env:

X_USERNAME="your_username"
X_EMAIL="your_email@example.com"
X_PASSWORD="your_password"

The script will attempt to log in and save cookies.json automatically. If X blocks the login (common), fall back to Option A.

Usage

GUI (recommended)

streamlit run gui.py

Opens a web dashboard at http://localhost:8501 with four tabs:

Tab	What it does
Token Lookup	Paste a CA, configure options, search — results shown in a sortable table
Watchlist	Browse all tracked callers with sorting, filtering, and expandable per-caller details
Update Outcomes	Check price performance for pending calls, view outcome breakdown
Settings	Import browser cookies, view file status and .env config

CLI

python solana_callers.py <CONTRACT_ADDRESS> [OPTIONS]

Arguments

Argument	Description	Default
`ca`	Solana token contract address	required*
`-o`, `--output`	CSV output file path	`callers.csv`
`--top`	Number of top callers to print	`5`
`--pump-time`	Manual pump timestamp (ISO 8601)	auto from DexScreener
`--update-outcomes`	Update outcome data for watchlist calls	—
`--user`	Only update outcomes for this username	all
`--force`	Re-check all outcomes, not just pending	—
`--batch-size`	Max outcome checks per run (0 = unlimited)	`0`
`--watchlist`	Print watchlist and exit	—
`--sort-by`	Sort watchlist: `tokens_called`, `win_rate`, `avg_multiple`, `total_score`	`tokens_called`
`--min-calls`	Filter: minimum completed calls	`0`
`--min-win-rate`	Filter: minimum win rate (0.0–1.0)	`0.0`

*Not required when using --update-outcomes or --watchlist.

Commands

Command	Description
`setup-cookies`	Import browser cookies for X authentication

Examples

# Basic — auto-detect pump time from DexScreener
python solana_callers.py ACtfUWtgvaXrQGNMiohTusi5jcx5RJf5zwu9aAxkpump

# Manual pump time (for delisted tokens)
python solana_callers.py 49bgrAJimCAGHG8o8VhSV7w1DhopLBrWutzYbqSFpump \
  --pump-time 2026-03-23T01:39:00Z

# Custom output and show top 10
python solana_callers.py ACtfUWtgvaXrQGNMiohTusi5jcx5RJf5zwu9aAxkpump \
  -o results.csv --top 10

# Investigate multiple tokens (watchlist builds up)
python solana_callers.py <TOKEN_1>
python solana_callers.py <TOKEN_2>
python solana_callers.py <TOKEN_3>

# Update outcomes for all pending calls
python solana_callers.py --update-outcomes

# Force re-check outcomes for a specific caller
python solana_callers.py --update-outcomes --user alpha_caller --force

# Batch update (useful for large watchlists)
python solana_callers.py --update-outcomes --batch-size 50

# View watchlist ranked by win rate, high performers only
python solana_callers.py --watchlist --sort-by win_rate --min-calls 5 --min-win-rate 0.5

Output

Console

Pair:          LOCKDOWN/SOL
DEX:           pumpfun
Created at:    2026-03-23T01:39:24+00:00
Est. peak:     2026-03-23T01:54:24+00:00
[twikit] Loaded saved cookies.

Searching X: ACtfUWtgvaXrQGNMiohTusi5jcx5RJf5zwu9aAxkpump
[twikit] Total raw tweets collected: 47
[twikit] Tweets in target window: 12

================================================================================
 Top 5 Influential Callers
================================================================================
 1. @alpha_caller        followers=82000     likes=312   RTs=88    score=33128.4
 2. @degen_scout         followers=45000     likes=201   RTs=44    score=18089.0
 3. @sol_sniper          followers=23000     likes=89    RTs=31    score=9241.8
 4. @memecoin_alerts     followers=15200     likes=44    RTs=12    score=6100.0
 5. @pump_watcher        followers=4410      likes=3     RTs=0     score=1765.2

==========================================================================================
 Watchlist — Top 5 Callers (sorted by win_rate)
==========================================================================================
 1. @alpha_caller        tokens=3   WR=67%  avg=4.2x  (3 rated)
 2. @degen_scout         tokens=2   WR=50%  avg=3.1x  (2 rated)
 3. @sol_sniper          tokens=2   WR=50%  avg=2.8x  (2 rated)

CSV (`callers.csv`)

Column	Description
`username`	X handle
`followers`	Follower count at time of tweet
`tweet_time`	ISO 8601 timestamp
`likes`	Like count
`retweets`	Retweet count
`score`	Influence score
`tweet_url`	Direct link to tweet

Watchlist (`watchlist.json`)

Persists across every run. Per caller:

{
  "alpha_caller": {
    "tokens_called": 3,
    "completed_calls": 3,
    "pending_calls": 0,
    "wins": 1,
    "big_wins": 1,
    "moonshots": 0,
    "losses": 1,
    "rugs": 0,
    "win_rate": 0.6667,
    "avg_multiple_24h": 4.2,
    "median_multiple_24h": 3.8,
    "best_call": { "ca": "ACtf...", "multiple": 8.2, "tweet_url": "..." },
    "worst_call": { "ca": "49bg...", "multiple": 0.4, "tweet_url": "..." },
    "first_seen": "2026-03-20T14:22:00+00:00",
    "last_seen": "2026-04-01T09:15:00+00:00",
    "calls": [
      {
        "ca": "ACtfUWtgva...",
        "tweet_time": "2026-03-23T01:35:00+00:00",
        "tweet_url": "https://x.com/alpha_caller/status/...",
        "score": 33128.4,
        "followers": 82000,
        "likes": 312,
        "retweets": 88,
        "outcome": {
          "status": "big_win",
          "entry_price": 0.0000234,
          "checked_at": "2026-03-24T01:40:00+00:00",
          "windows": {
            "15m":  { "multiple": 1.4, "drawdown": -0.12, "final": 0.0000287 },
            "1h":   { "multiple": 2.1, "drawdown": -0.18, "final": 0.0000412 },
            "6h":   { "multiple": 3.8, "drawdown": -0.18, "final": 0.0000356 },
            "24h":  { "multiple": 8.2, "drawdown": -0.31, "final": 0.0000189 },
            "7d":   { "multiple": 8.2, "drawdown": -0.78, "final": 0.0000051 }
          },
          "ath": { "multiple": 8.2, "minutes_to_ath": 187 }
        }
      }
    ]
  }
}

The watchlist deduplicates by tweet URL, so re-running the same token is safe.

Outcome Labels

Label	Threshold
`moonshot`	Best 24h multiple >= 10x
`big_win`	Best 24h multiple >= 5x
`win`	Best 24h multiple >= 2x
`loss`	Best 24h multiple < 1.5x
`rug`	Multiple < 1x AND drawdown > 50% in 1h
`pending`	Less than 24h since call
`no_data`	No candle data available

Scoring Formula

score = (followers × 0.4) + (likes × 0.4) + (retweets × 0.2)

Only tweets posted before the estimated peak are scored. If no pre-peak tweets are found (e.g. X's search index dropped them), all returned tweets are scored instead.

Troubleshooting

Problem	Solution
`Couldn't get KEY_BYTE indices`	The script patches this automatically. Make sure you're on the latest version (`git pull`).
`401 Unauthorized`	Cookies expired. Re-run `python solana_callers.py setup-cookies`.
`All search query IDs returned 404`	X rotated GraphQL endpoints. Get the current ID from browser DevTools (Network → filter `SearchTimeline`), then set `TWITTER_SEARCH_QID=<id>` in `.env`.
`No pairs found on DexScreener`	Token was delisted. Use `--pump-time` to set the timestamp manually.
`No tweets found`	X's search index doesn't reliably cover long random strings (CAs) or old tweets. Try a more recent token.
`ModuleNotFoundError: aiohttp`	Run `pip install -r requirements.txt` inside your virtual environment.
`git pull` conflicts	Run `git stash && git pull && git stash pop`, or `git checkout -- solana_callers.py && git pull` to discard local changes.

Environment Variables

All optional. Set in .env or export in your shell.

Variable	Purpose	Default
`X_USERNAME`	X username (login fallback)	—
`X_EMAIL`	X email (login fallback)	—
`X_PASSWORD`	X password (login fallback)	—
`TWITTER_SEARCH_QID`	Override GraphQL search query ID when X rotates endpoints	—
`OUTCOME_CHECK_ENABLED`	Auto-check outcomes after each CA lookup	`true`
`OUTCOME_CHECK_DELAY_MIN`	Minutes to wait before checking a call's outcome	`60`
`OUTCOME_CHECK_BACKOFF_HOURS`	Hours between re-checks for pending calls	`24`

Files

File	Tracked	Description
`solana_callers.py`	Yes	Main script — CA lookup, scoring, watchlist
`outcomes.py`	Yes	Outcome tracking — candle fetching, classification, aggregates
`gui.py`	Yes	Streamlit web GUI
`requirements.txt`	Yes	Python dependencies
`.env.example`	Yes	Template for credentials and config
`.env`	No	Your credentials (gitignored)
`cookies.json`	No	X session cookies (gitignored)
`watchlist.json`	No	Persistent caller database with outcomes (gitignored)
`callers.csv`	No	Output from last run

Name		Name	Last commit message	Last commit date
Latest commit History 23 Commits
.claude		.claude
specs		specs
.env.example		.env.example
.gitignore		.gitignore
README.md		README.md
callers.csv		callers.csv
config.py		config.py
gui.py		gui.py
outcomes.py		outcomes.py
outcomes_debug.log		outcomes_debug.log
requirements.txt		requirements.txt
solana_callers.py		solana_callers.py

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

Solana Token Early Caller Finder

How It Works

Requirements

Installation

Dependencies

Setup — X Authentication

Option A — Browser cookies (recommended)

Option B — Credentials via .env

Usage

GUI (recommended)

CLI

Arguments

Commands

Examples

Output

Console

CSV (`callers.csv`)

Watchlist (`watchlist.json`)

Outcome Labels

Scoring Formula

Troubleshooting

Environment Variables

Files

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

Solana Token Early Caller Finder

How It Works

Requirements

Installation

Dependencies

Setup — X Authentication

Option A — Browser cookies (recommended)

Option B — Credentials via .env

Usage

GUI (recommended)

CLI

Arguments

Commands

Examples

Output

Console

CSV (callers.csv)

Watchlist (watchlist.json)

Outcome Labels

Scoring Formula

Troubleshooting

Environment Variables

Files

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

CSV (`callers.csv`)

Watchlist (`watchlist.json`)

Packages